Reference

Benchmarks

Every run loads the same randomized data into a single Logs process on MinIO, next to Loki, then sends both the same randomized LogQL queries. Each answer is checked for correctness and timed. The newest run is shown by default.

Runhistoricalvs Loki 3.5.5s3Full report ↗

Failed· 2 mismatch case(s)2026-10-08 20:32Z · 0b54fe965a (dirty) · simd-perf · 30 min · OrbStack · arm64 · 12 CPUs · 8.8 GiB

9,154

queries compared

mismatches

353 non-match total

10.1 ms

p50 query

11.9 ms Loki

252 ms

p99 query

539 ms Loki

0.54×

p50 latency ratio

per query, impl / oracle

0.23

mean CPU cores

0.40 Loki

Latency by percentile

Plural LogsLoki 3.5.5

Latency of every query both sides answered. Each query runs against both systems with the same data, so the percentiles compare like with like.

By query family

Query family	Cases	p99 latency	p50 ratio	Non-match
logql.metric.range	3,231	341 ms 1.1 s	0.47×	154
logql.metric.instant	1,922	151 ms 195 ms	0.67×	88
logql.log.parsed	1,497	191 ms 402 ms	0.50×	17
logql.log	1,191	179 ms 385 ms	0.60×	26
logql.log.parsed.limited	348	126 ms 120 ms	0.67×	7
logql.labels	341	39.8 ms 28.2 ms	0.49×	0
logql.label_values	314	40.1 ms 28.6 ms	0.55×	47
logql.log.limited	310	210 ms 115 ms	0.86×	14

Bars share one linear scale across families. The ratio is the median of per-query implementation / oracle latency, so values below 1× mean Plural answered faster.

Correctness

match8,801
both error223
inconclusive69
oracle unstable47
oracle error6
oracle timeout6
mismatch2

Inconclusive results are differences the API contract allows, such as both sides truncating at a limit.

Resources

CPU, mean cores

Plural

0.23

Loki

0.40

Memory, p95 working set

Plural

743 MiB

Loki

695 MiB

Sampled from the Docker Engine API every two seconds over the run. Object storage is excluded from both sides.

History · historical · Loki 3.5.5

Plural LogsLoki 3.5.5

Query latency for every run on this scenario and oracle. Click a point to open that run.

How runs are measured

The differential fuzzer in tests/regression/harness/fuzz runs both systems on one Docker host, each pinned to its own CPU set. A run fails on any mismatch, implementation error or timeout, unstable implementation answer, or rejected write. The recent scenario queries data still being ingested; historical queries older data that has been flushed to object storage, with artificial latency added to MinIO.

To record a new run from the repository root:

bash

PYTHONPATH=tests/regression mise exec -- python -m harness.fuzz.bench --products logs --duration 30m

Results land in documentation/benchmarks/fuzz and appear here on the next docs build.